Abstract

Structuralism is often associated with a program, in keeping with the Durkheimian tradition, of reducing social norms to a kind of causality. On this reading, Émile Durkheim's collective representations became, in Claude Lévi-Strauss' work, cognitive or logical constraints. If so, then structuralism falls under Wittgenstein's objections to treating rules as causes. What this article shows, however, is that this reading of structuralism is misguided. The necessity and justification of introducing structural methods, first in linguistics and then in anthropology, as well as the general concept of the sign, were rooted in the perception of a problem too often overlooked: that of the characterization of the data in the social and cultural sciences. Just as Saussure showed that the real problem of linguistics is not, what does it mean? but what are the units of language?, so Lévi-Strauss showed that social sciences should not be concerned with the causes of human behaviors but instead with the categorization of actions. In both cases, it appears that an individual action is the actualization of a common practice, habit, or custom that, itself, cannot be reduced to a set of observable features but must be defined by its differential position in a system. Structural anthropology thus leads to a truly ontological problem, which resonates interestingly today, in the context of the “ontological turn.”